The Internet, Social Capital, and Subjective Well-Being Fabio Sabatini Sapienza University of Rome Third International Annual Conference of the LCSR Moscow, Russia November 12-17, 2013
• In the years that preceded the social networking revolution, indicators of social participation have declined in many OECD countries (Bartolini et al., 2013; Costa & Kahn, 2003; Putnam, 2002; Sarracino, • The project was 2010). motivated by two stylised facts : • However, more recently, the success of social networking sites (SNSs) has resulted in a steep rise in online social participation (Antoci et al. 2012a; 2012b; 2013; Brenner and Smith, 2013).
Some facts • As of May 2013, according o Pew Research Center (PRC) Internet & American Life Project Survey data: • 72% of U.S. online adults were active on SNSs (67% use Facebook, 16% use Twitter, 15% use Pinterest and 13% use Instagram). • 80% of online young adults (aged 18 – 29) use SNSs • 77% of middle-aged adults (30 – 49) use SNSs • 50% create and share original content online. • 46% of adult users post original photos or videos online that they themselves have created.
Main research questions • In the “social networking era”, may Internet use accelerate the decline in social participation documented by empirical studies? • Or does it offer a way to support social relationships against the threats posed by the disruption of ties and the weakening of community life? • How does social networking affect trust? • Which is the overall impact on social and individual well-being?
Preliminary research outcomes • The development of the project has so far led to the preparation of two papers respectively describing a theoretical and an empirical analysis. • Theoretical paper : Antoci, A., Sabatini, F., Sodini, M. (2013). Social poverty traps and Internet-mediated interactions. Can social networks save human relations? • Empirical paper : Sabatini, F., Sarracino, F. (2013). Will Facebook save or destroy social capital? An empirical investigation into the role of online interactions in social trust and networks. • Printed copies of these papers have been distributed to participants.
Will Facebook save or destroy social capital? An empirical investigation into the role of online interactions in the evolution of trust and networks Fabio Sabatini Francesco Sarracino
Pessimistic views on Internet use • The more time people spend using the Internet during leisure time, the more time has to be detracted from social activities. 3 • The Internet allows users to conduct many daily transactions such as shopping or banking online from home: this may reduce main arguments : the opportunities for face to face interactions. • “Community without propinquity” argument (Webber, 1963). A heterogeneous urban environment would be characteristic of the absence of “intimate personal acquaintanceship” and would result in the “segmentation of human relations” into those that were “largely anonymous, superficial , and transitory” (Wirth, 1938, p. 1).
Pessimistic views on Internet use • These studies were in most cases conducted before the online networking revolution, when Internet use mainly was an individual activity. 3 • Small and biased samples (usually, limited main weaknesses : communities of undergrad students). • Endogeneity issues not always completely addressed.
Opimistic views • A couple of pioneer economic studies support the intuition that Internet use may actually not discourage sociability and participation (Pénard and Poussing, 2010; Bauernschuster et al., 2011). • Empirical studies in the fields of applied psychology and communication science have specifically analyzed how online networking influences social interactions across Internet users (e.g. many papers in the Journal of Applied Social Psychology and Journal of Computer Mediated Communication ). Most of these studies find that networking supports social interactions, trust, self-esteem and sbw in small (and non representative!) samples .
A collection of papers in this field can be browsed – and downloaded – from the Social Capital Gateway at the url: www.socialcapitalgateway.org/internet
Contribution to the literature • We add to this multidisciplinary debate by carrying out the first empirical study on the effect of online interactions through: 1) social networking sites (SNSs) 2) chats, newsgroups, and forums on two dimensions of social capital: 1) trust towards unknown others (hereafter “social trust”) – cognitive SC 2) social networks developed through face-to- face interactions among friends and acquaintances – structural SC in a large and representative sample of the Italian population ( n = 117,878).
Social capital • The structural dimension is given by face to face interactions, as measured by the frequency of meetings with friends . Respondents were asked to report how many times they meet their friends on a scale from 1 (in case they have no friends) to 7 (if respondents meet their friends everyday).
Social capital • The Rosenberg question – “Generally speaking…” • Cognitive dimension - Social trust was • The “wallet question”: In the city or measured through: area where you live, imagine you lost your wallet holding money and your identification or address and it was found by someone else. How likely do you think your wallet would be returned to you if it were found by a stranger?”.
Online networking • We measure social networking with two dichotomous variables reporting: 1) Participation in social networking sites such as Facebook and Twitter. 2) Participation in chats, newsgroups, and forums.
Identification issues Common bias : Individual effects, such as personal characteristics or exogenous shocks, may be correlated with both online networking and the two dimensions of social capital we account for. We cannot discard the hypothesis that online networking is endogenous to social Reverse causality might also interactions and arise . For example, people who social trust meet their friends frequently may be encouraged to join online networks to strengthen existing social ties. Reverse causality may also work in the opposite direction to the extent to which people who have no (or just a few) friends may look for interactions on Facebook to alleviate their social isolation.
Identification strategy • Large set of covariates • Identification strategy based on: • Instrumental variables
Instruments 1) The percentage of the population for whom a DSL connection was available in respondents’ region of residence. 2) Regional digital divide, given by the percentage of the region’s area that was not covered by optical fibre . • Both the instrumental variables were measured in 2008 , two years before the first wave of the Multipurpose Household Survey we employ in our study.
Instruments’ orthogonality • 2008 level of regional DSL or fibre coverage cannot per se exert a direct influence on individual social capital. Rather, their availability in the area creates the premise for the individual choice to purchase a fast-speed access and, subsequently, to participate in SNSs. • DSL or fibre coverage in the region of residence have never been found to be correlated with social interactions and social trust at the individual level. • DSL and fibre coverage in 2008 cannot be endogenous – in the sense of reverse causality – to the individual involvement in online networks in 2010-2012. • The digital divide in terms of fibre coverage also depend on orographic differences between regions, which generated a variation in access to fast Internet across Italian regions that is exogeneous to people’s social capital. • The assumption of orthogonality of the instruments is not disconfirmed by the tests of over-identifying restrictions we run in the context of IV estimates.
Controls • The technology that respondents used to connect to the Internet. Possible categories were cable broadband (optical fibre, intranet, PLC, etc.), satellite or other wireless connections (e.g. wi-fi and wi-max), wireless connection through a tablet and or a mobile phone employing a 3G mobile telecommunication technology, wireless connection employing a 3G modem (e.g. a USB key), or connection with a WAP or a GPRS mobile phone. • Age (both in linear and squared form), gender, marital status, the number of childen, education, work status, and • The time spent in commuting (in minutes).
Main results • Face to face interactions are strongly significantly and POSITIVELY correlated with online networking . • Social trust – however measured – is significantly and NEGATIVELY correlated with online networking . These results are robust: 1) to all the estimation methods we employed in the analysis, including probit and ordered probit estimates, 2SLS estimates conducted through the CMP estimator developed by Roodman (2009), and other techniques. 2) to all the models specification we tested.
Other results • Women show: 1) a significantly lower propensity for participation in networks like Facebook and Twitter and in chats, forums, and newsgroups. 2) a significantly lower propensity for face to face interactions. 3) a significantly higher propensity to trust others.
Prob of meeting friends less than 4 times per year
Prob of meeting friends EVERYDAY
Prob of meeting friends very frequently
Recommend
More recommend
